Depvartmient of Poultry Research, Ifye College (University of London), Ashford, Kent

J. Anat. (1965), 99, 3, pp. 485-506 485 With 18 figures Prin#d in Great Brtiain The blood supply to the avian oviduct, with special reference to the shell gland BY R. D. HODGES Depvartmient of Poultry Research, IFye College (University of London), Ashford, Kent

INTRO D U CT ION Until recently there was no detailed account of the blood supply to the avian shell gland. This situation was partially rectified by Freedman (1962) and Freedman & Sturkie (1963). Previous to this the only traceable accounts of the blood supply to the oviduct as a whole were those of Barkow (1829), Neugebauer (1845), Owen (1866), Streseman (1928), Mauger (1941) and Bradley & Grahame (1960), and to the shell gland in particular, those of Neugebauer (1845), Streseman (1928) and Hun- saker (1959). None of these accounts dealt with the subject in any detail. The shell gland withdraws the main component of the shell, calcium carbonate, from its blood supply in the form of calcium and carbonate ions at the same time that the shell is being laid down; there being no storage of calcium in the gland (Richardson, 1935). Consequently a detailed understanding of its vasculature is essential to any physiological investigations into the mechanism of shell secretion. The following account is an attempt to produce a detailed comparative study of the blood supply to the oviduct, and in particular the shell gland, of the three most common domesticated birds, the fowl, the turkey and the duck.

MATERIALS AND METHODS The fowls (fifty-eight birds) used in this study were actively laying Light Sussex hens in the first or second laying season; the ducks (six birds) were mature breeding stock of the Aylesbury strain, and the turkeys (two birds) were actively laying Beltsville White hens. The majority of the birds, forty-nine fowls, the six ducks and the two turkeys, were injected with suspensions of red and blue neoprene latex (Neoprene Latex no. 572, Messrs Bostik Ltd., Leicester) in order to demonstrate the macroscopic and microscopic anatomy. The remaining nine fowls were injected with undiluted indian ink (Reeves waterproof indian ink) to demonstrate only the microscopic anatomy. The latex injection technique was as follows. The bird was killed with an intra- venous injection of pentabarbitone and the breast bone and liver removed. A poly- vinyl cannula was tied into the aorta below the heart and the blood was flushed out of the abdominal region with physiological saline warmed to 40° C., using a pressure of 2 ft. of water augmented by a manual pumping action on the supply tube. The latex was injected by means of an adaptation of the apparatus described by Tompsett (1956). A rubber hand bellows with a Winchester bottle reservoir and a mercury manometer were used as a pressure source in place of Tompsett's oxygen cylinder. The latex was injected at approximately normal blood pressures, i.e. 486 R. D. HODGES 1:30 mm. of mercury for the arteries and about 10 mm. of mercury for the veins. The arteries were injected with red latex through the aortic cannula, and the veinskere injected with blue latex by means of a cannula tied in the posterior vena cava. The latex coagulated in situ within an hour or so of injection. The latex casts of the blood vessels were obtained by corroding shell glands in concentrated hydrochloric acid for 72 hr. and then washing with water. Indian ink injections were made in the same way through the aorta, the ink passing through the capillary bed of the shell gland to fill the veins. All specimens were fixed and stored in 10 % formol saline. Portions of the ink-injected glands were examined in sections stained with Ehrlich's haematoxylin and eosin or by clearing with benzene and benzyl benzoate (Orsini, 1962). In both turkeys and three of the ducks the shell glands contained eggs. In the fowls forty-five glands contained eggs whilst the remaining thirteen were in the resting phase. Of the glands which contained eggs no attempt was made to differentiate the stages of shell formation. Although forty-nine of the fowls were injected with latex, in the following account the numbers cited frequently do not add up to this total, due partly to incomplete injections in some earlier cases and partly to vascular abnormalities in certain cases. The folds of the oviduct overlapped each other considerably, and in order to examine the blood supply closely it was necessary to cut the mesenteries. The diagrams represent the oviducts in a partially unfolded position; consequently manv of the vessels are elongated.

NO MEN C LATURE Most of the nomenclature adopted in this account is that suggested by Freedman (1962) and Freedman & Sturkie (1963). Where the present author differs from these authors on points of nomenclature, the subject is discussed. Table 1. Explanation of newly named blood vessels Vessel Origin Distribution or drainage Anterior oviducal artery and The ovarian vessels Infundibulum and anterior vein magnum Middle uterine artery Femoral artery Most of magnum Middle uterine veins (except Renal rehevens vein Infundibulum, magnum and hypogastric) isthmus Inferior oviducal artery and vein Numerous points of Ventral surface of oviduct, origin (see text) excluding shell gland and vagina Superior oviducal artery and Numerous points of Dorsal surface of infundibulum, vein origin magnum and isthmus Utero-vaginal artery and vein Posterior uterine artery Posterior areas of shell gland and and vein utero-vaginal junction Vaginal arteries and veins Left pelvic artery or vein MNIiddle and posterior regions of the vagina Some previously undesignated vessels, particularly oviduct vessels, have been named. These, together with their origin and termination, are listed in Table 1. In order that the nomenclature shall correspond as closely as possible with that Blood supply of the avian oviduct 487 of Freedman & Sturkie (1963) all shell gland vessels will be termed 'uterine vessels'. However, throughout this account the terms shell gland and uterus are synonymous. Frequent mention has been made of the lateral and medial surfaces of the shell gland. When describing vessels of a particular surface the full designation is often complicated, e.g. the lateral lateral uterine artery. In these circumstances the shell gland surface will be given first and with a capital letter, thus the Lateral lateral uterine artery. RESULTS The blood supply to the oviduct Figure 1 is a diagram of the main blood vessels of the left side of the fowl's abdomen and the branches of these vessels which supply the oviduct.

/ I

V - > ~~~~~;tiduct

Cloaca)

Fig. 1. Semi-diagrammatic representation of a ventral view of the abdominal blood vessels of the fowl and the branches supplying the oviduct and ovary. The oviduct has been displaced to the right making the oviduct vessels longer than normal. The kidney is shown in outline only. A, Anterior oviducal artery; B, hypogastric artery; C, left internal iliac artery; D, middle oviducal artery; E, anterior oviducal veins; F, middle oviducal veins; G, hypogastric vein; H, left internal iliac vein; J, inferior oviducal artery and vein; K, inferior uterine artery and vein; L, branches of anterior oviducal artery supplying inferior oviducal artery; M, lateral anterior uterine artery and vein; N, superior oviducal artery and vein; 0, dorsal aorta; P, renolumbar artery; Q, femoral artery; R, sciatic artery; S, ovarian artery; T, ovarian veins; U, renal rehevens vein; V, renal portal vein; W, pudendal artery and vein; X, posterior uterine artery and vein; Y, vaginal artery and vein; Z, pelvic artery and vein; DM, dorsal mesentery; ViM, ventral mesentery. 488 R. D. HODGES The blood supply to the oviduct in the fowl, turkey and duck is shown diagrammatically in Figs. 2-4 respectively. Variations from the normal pattern of vessels in the fowl are shown in Fig. 5. The blood supply was found to be very similar in all three species.

w Infundibulum ± Magnum i Isthmus w Shell gland w Vagina Fig. 2. Blood supply to the fowl's oviduct; normal pattern of vessels. Key as in Fig. 1. In Figs. 2-5 the method of representation of the blood vessels is as follows. The cross- hatched vessels are unaccompanied arteries; the black vessels are veins unaccompanied by arteries; and the white vessels represent arteries and veins running in parallel. Thus when an artery and vein meet, if the vessels passing away from the junction are white, it indicates that both the original vessels have divided to send parallel branches along the white vessels (Fig. 2, B, F). If, however, the artery and vein continue in their original state after the junction (Fig. 5, D, F) then this indicates merely a crossing over of vessels.

Fig. 3. Blood supply to the turkey's oviduct; normal pattern of vessels. Key as in Fig. 1. The fowl Branches of the ovarian, hypogastric, femoral and left internal iliac arteries supplied the oviduct as follows: (1) A branch of the ovarian artery left the ovary and, coursing posteriorly along the surface of the anterior lobe of the left kidney, passed into the dorsal mesentery Blood supply of the avian oviduct 489 where it divided up to supply the infundibulum and anterior magnum (Fig. 2, A). This artery was termed the anterior oviducal artery. (2) In two out of the ten cases a branch of the left femoral artery passed between the anterior and middle lobes of the kidney and divided up to supply the middle and posterior magnum (Fig. 5, D). Although atypical, this vessel has been described by 'Mauger (1941) and was designated the middle oviducal artery.

D F

Fig. 4. Blood supply to the duck's oviduct; normal pattern of vessels. Key as in Fig. 1.

Fig. 5. Blood supply to the fowl's oviduct; variations on the normal pattern of vessels. Key as in Fig. 1. (.3) The hypogastric artery (Fig. 2, B) arose as a branch of the left sciatic artery between the middle and posterior renal arteries. It passed ventro-posteriorly through the upper mesentery until it impinged upon the dorsal surface of the shell gland. Between its origin and this point the artery supplied from three to six branches to the posterior magnum and the isthmus. (4) The left internal iliac artery (Fig. 2, C) supplied blood to the posterior shell gland and the vagina through its middle uterine, posterior uterine and vaginal branches. The venous drainage occurred through the ovarian, renal rehevens and internal iliac veins, thus: (1) A vein or occasionally two small veins which passed into the ovarian venous 490 R. ID. HODGES complex (Fig. 2, E) draining a similar area to that supplied by the anterior oviducal artery and therefore termed the anterior oviducal vein or veins. In some cases this vein only drained the infundibular funnel, the infundibular neck and anterior magnum regions draining into the renal rehevens vein. In one case this vein was absent (Fig. 5). (2) A number of veins, between three and six (Fig. 2, F), which drained into the renal relievens vein, the most anterior entering the rehevens vein on the anterior lobe of the left kidney and the most posterior, the hypogastric vein (Fig. 2, G), entering on the posterior lobe. These veins were irregular in number, size and position and it was impossible to name them separately, apart from the hypogastric vein which was always the most posterior and always drained the anterior shell gland and posterior isthmus. However, in order that these vessels, excluding the hypogastric vein, should have some form of designation they were grouped as the middle oviducal veins. (3) The left internal iliac vein (Fig. 2, H) drained the posterior shell gland and vagina through middle, superior and posterior uterine veins and vaginal veins. The inferior oviducal artery and vein were well-developed vessels which ran almost the full length of the oviduct in the lower mesentery (Fig. 2, J). They were sometimes closely applied to the oviducal surface and sometimes separated from it. They began on the infundibular funnel and continued along the ventral surface of the magnum and isthmus on to the shell gland (as the inferior uterine vessels, Fig. 2, K), and occasionally on to the vagina as well. The inferior oviducal artery was supplied by one or two branches of the anterior oviducal artery running down the lateral face of the oviduct in the region of the anterior magnum (Fig. 2, L), by the Lateral anterior uterine artery (Fig. 2,111) and also by a variable number of small branches of the anterior and middle oviducal arteries and of the hypogastric artery (Figs. 2 and 5). The superior oviducal artery and vein ran, in a similar manner to the inferior vessels, along the dorsal surface of the oviduct (Fig. 2, N). However, they were not so well developed as the inferior vessels and often did not follow a continuous course along the dorsal surface of the infundibulum, magnum and isthmus. Throughout the whole vascular tree to the oviduct there were frequent arterial and venous anastomoses at all levels. The turkey The only major difference between the fowl and the turkey was the absence of the anterior oViducal vessels from the ovarian blood supply, the anterior end of the oviduct being supplied from the femoral artery through the middle oviducal artery and drained into the anterior end of the rehevens vein (Fig. 3, D, F). Otherwise there was a similar arterial supply from the hypogastric and left internal iliac arteries and a venous drainage through the left internal iliac vein, the hvpogastric vein and a number, usually 4, of middle ox iducal veins. The duck The differences between the fowl and the duck were as follows (Fig. 4): a small anterior oviducal artery was found in only one duck; no anterior oviducal veins Blood supply of the avian oviduct 491 were found. Posterior uterine veins were absent in all except one case where there was a very small Lateral posterior uterine vein. There were only from two to four venous outlets into the rehevens vein and the number of branches of these veins and of the middle oviducal and hypogastric arteries was so small that the superior oviducal artery and vein were better developed than in the fowl and turkey.

The blood supply to the shell gland The fowl Arteries of the lateral side Anterior uterine artery. The hypogastric artery impinged upon the surface of the shell gland about one quarter of the way back along the dorsal surface (Fig. 6, 1). As it did so it normally divided up into two or more branches. These were usually Lateral and Medial anterior uterine arteries, and often a superior uterine artery. The lateral anterior uterine artery normally passed vertically down the lateral face of the gland (Fig. 6, 2). From the anterior edge of the artery arose a number of small branches which supplied the area of the gland between the artery and the isthmo-uterine junction. These arteries often anastomosed with arteries supplying the posterior isthmus. When the anterior uterine artery reached the ventral surface of the gland it divided into two, one branch, the inferior oviducal artery, passing forward along the ventral surface of the isthmus, the other branch passing posteriorly on to the ventral surface of the gland as the inferior uterine artery (Fig. 6, 3 and 4). From the posterior edge of the anterior uterine artery there arose small branches supplying the area just posterior to the artery, and occasionally the area of the gland under- lying the superior uterine artery. Inferior uterine artery. The artery began at the ventral division of the anterior uterine artery and followed a rather tortuous path posteriorly along the ventral surface of the shell gland. When strongly developed it supplied the ventral surface of the organ by means of small branches. At the same time branches from the Lateral lateral uterine artery and the medial lateral uterine artery passed ventrally to anastomose with these branches of the inferior uterine artery. The posterior end of the artery normally passed on to either the lateral or medial sides of the gland, anastomosing with a branch (Fig. 6, 5) of the corresponding posterior uterine artery. However, occasionally the inferior uterine artery divided to fuse with branches of the posterior uterine arteries of both sides (eight cases). When strongly developed the inferior uterine artery often sent branches on to the ventral surface of the vagina (fourteen cases, Fig. 6, 6), and occasionally, instead of terminating as described above, it passed on to the ventral vagina to anastomose with the vaginal arteries (six cases). When poorly developed the artery either passed only part way along the ventral surface of the shell gland or more frequently it occurred as a series of fine arteries following an erratic course between well developed vessels passing round the ventral surface between the lateral uterine arteries (Fig. 7, 4). Lateral uterine artery. The origin of this artery was normally in one of two sites. Most frequently (twenty-four out of forty-six cases) it occurred as a branch of the 31 Anat. 99 492 R. D. HODGES superior uterine artery (Fig. 6, 7). However, in eighteen cases it arose from the posterior edge of the anterior uterine artery (Fig. 7, 7). In the first case, where the artery originated from the superior uterine artery, it passed centrally and, at the approximate mid-point of the lateral side, it then turned and passed posteriorly (Fig. 6, 7). At this point the downward direction was often continued as a medium-sized artery which nearly always connected with the inferior uterine artery. Where the superior uterine artery was short and the vertical continuation of the lateral uterine artery was strongly developed it gave the appear- ance of an aberrant second anterior uterine artery (two cases). Occasionally, where the superior uterine artery was not present and the lateral uterine artery originated from the hypogastric artery, the lateral artery passed diagonally downwards and backwards from its origin (four cases). In the second case, where the lateral uterine artery occurred as a branch of the anterior uterine artery, it passed horizontally backwards to its point of termination. The lateral uterine artery gave off well-developed branches which passed upwards and downwards to supply the central area of the lateral face of the gland. These branches were particularly well developed in cases where the inferior or superior uterine arteries were poorly developed (Fig. 7) and they then spread round on to the superior or inferior surfaces of the gland. In most cases the lateral uterine artery appeared to terminate at the point where it merged with the posterior uterine artery (Fig. 6, 9), i.e. at the point where the latter impinged upon the surface of the shell gland. However, in thirty-six cases an artery branched off from the region of the junction of the lateral and posterior uterine arteries, frequently giving the impression that the lateral uterine artery continued beyond the junction to supply areas of the posterior end of the gland. Branches of this artery (Fig. 6, 10) tended to pass across the utero-vaginal junction on to the vagina, and in twelve cases the artery anastomosed with the vaginal arteries (Fig. 7). Whenever it occurred this artery was well developed and thus would seem to warrant a special name. It is therefore proposed to term it the utero-vaginal artery. Where this artery did not occur the posterior end of the shell gland was supplied by a number of small branches arising from the posterio-lateral junction and from the artery which connected the inferior and posterior uterine arteries. There were seven cases in which no posterior uterine artery occurred and the lateral uterine artery did continue past its normal point of termination to supply all the posterior end of the gland. In five of these cases the artery eventually anastomosed with the vaginal arteries; in the sixth case it passed posteriorly to merge with the posterior uterine artery on the medial side; in the seventh case it terminated on the utero-vaginal junction. Superior uterine artery. This artery (Fig. 6, 8) normally branched off from the hypogastric artery where the latter divided to form the anterior uterine arteries. It then passed posteriorly along the dorsal surface and terminated at the division of the two lateral uterine arteries. The size and length of this artery was extremely variable. Out of fortv-nine cases there were fifteen in which there was no superior uterine artery at all; in fourteen others it was very short, dividing almost immedi- ately into lateral uterine arteries. In twelve cases it was about a quarter of the Blood supply of the avian ovidu( 49t493 overall length of the gland; in four it was about half the length, and in only one case was it over half the length of the gland. In two cases a small continuation of the superior uterine artery occurred which anastomosed with the posterior uterine artery. Posterior uterine artery. This artery (Fig. 6, 9) had a rather variable point of origin and point of impinging upon the gland. When these points were at the anterior end of the range of variation, that is where the artery branched off from the internal iliac artery where the latter divided to form the pelvic and pudendal arteries, and where it entered the substance of the gland at a point about two-thirds of the distance along the length of the gland, the artery could be considered to be a

3 4 6~~~~~~~~~~~1

± Isthmus 1 Shell gland 1 Vagina 1 Fig. 6. Lateral side of the fowl's shell gland; normal pattern of arteries. 1, hypogastric artery; 2, anterior uterine artery; 3, inferior oviducal artery; 4, inferior uterine artery; 5, artery connecting inferior uterine artery with posterior uterine artery; 6. vaginal branch of inferior uterine artery; 7, lateral uterine artery; 8, superior uterine artery; 9, Lateral posterior uterine artery; 10, utero-vaginal artery; 11, point of junction of the lateral and posterior uterine arteries; 12, Medial posterior uterine artery; 13, middle uterine artery; 14, left internal iliac artery; 15, left pudendal artery; 16, left pelvic artery; 17, vaginal arteries; 18, superior oviducal artery. middle uterine artery. But as these points came closer to the utero-vaginal junction, so it became more plainly a posterior uterine artery. Due to this range of variation all these arteries have been termed posterior uterine arteries, and the term middle uterine artery has been reserved for a vessel to be described below. This nomenclature is at variance with that of Freedman & Sturkie (1963). Several branches arose from the point of junction of the posterior uterine artery and the lateral uterine artery (Fig. 6, 11). First, there were one or two arteries which passed downwards to fuse with the inferior uterine artery (Fig. 6, 5). Secondly, there was the utero-vaginal artery. In all but seven of the cases there were two posterior uterine arteries, one on 31-2 494 R. D. HODGES either side. In all of the seven there was a Medial posterior uterine artery, the Lateral artery being the missing one. It was replaced by a branch of the pelvic artery which supplied the vagina, the lateral uterine artery continuing on to the vagina to anastomose with it. 14

Fig. 7. Lateral side of the fowl's shell gland; variations on the arterial pattern. Key as for Fig. 6.

3 It Fig. 8. Lateral side of the turkey's shell gland; normal arterial pattern. Key as for Fig. 6.

The Lateral posterior uterine artery almost always branched off from the pelvic artery before the Medial posterior uterine artery (Fig. 6, 12), and this was corres- pondingly reflected in the position of the anastomosis with the lateral uterine artery, the artery with a more posterior point of origin having a more posterior point of anastomosis. After the pelvic artery had given off the posterior uterine arteries it continued posteriorly until it reached the vagina. Here it divided to give one or more branches Blood supply of the avian oviduct 495 supplying each side of the vagina, the vaginal arteries, and in some cases, nine out of forty-nine, continued on to supply the cloaca (Fig. 7). Aliddle uterine artery. This artery did not occur very frequently (sixteen cases only). In only two cases did it supply the lateral side and in only one case was there both Lateral and Medial middle uterine arteries present. The origin of the middle uterine artery was very variable. It either arose from the left pelvic artery at, or caudal to, its bifurcation with the pudendal (Fig. 6, 13); or it arose as a branch of the Lateral or Medial posterior uterine arteries after these had left the pelvic artery. It then passed anteriorly before dividing up to supply the shell gland. In most cases this artery was poorly developed. However, in seven on the medial side and two on the lateral, it was well developed and eventually anastomosed with the lateral uterine artery on the corresponding side (Fig. 10, 13).

2~~ ~ 15

Fig. 9. Lateral side of the duck's shell gland; normal arterial pattern. Key as for Fig. 6. The turkey The arterial supply of the shell gland in the turkey was similar to that of the fowl. The differences were as follows. The superior uterine artery was well developed; it traversed most of the length of the dorsal surface and in one case it fused posteriorly with the pelvic artery. Branches of the superior uterine artery supplied the dorsal regions of both lateral and medial sides. Two of these branches coming off at about the mid-point of the dorsal surface eventually joined up with the posterior uterine arteries of each side and are thus termed the lateral uterine arteries (Fig. 8, 7) as they more or less corresponded with those in the fowl. They were, however, no larger than the other branches and were not so well developed as in the fowl. The duck The inferior uterine artery gave off branches to the lateral and medial sides and divided posteriorly into three branches, one passing on to the ventral vagina and one passing up each side of the gland to fuse with the posterior uterine arteries 496 R. D. HODGES (Fig. 9). The lateral uterine arteries resembled those of the turkey in that they ran more along the superior surface of the shell gland than along the lateral or medial surfaces. They arose from a branch of the hypogastric artery and merged with the posterior uterine arteries. The lateral and medial surfaces were supplied by small arteries passing vertically between the inferior and lateral uterine arteries. There was no true superior uterine artery. The hypogastric artery gave off one or two major branches before it impinged upon the surface of the shell gland and these, sometimes giving off secondary anterior uterine arteries, became the lateral uterine arteries when they contacted the wall of the shell gland (Fig. 9). The posterior uterine arteries normally arose from a single branch of the pelvic artery and impinged upon the gland superiorly close to the utero-vaginal junction. They gave off one or two small branches to the junction and then fused with the lateral uterine arteries. The utero-vaginal artery, if present, was not well developed. Apart from these instances the blood vessel pattern was basically similar to that found in the fowl. Thefowl Arteries of the medial side The arteries of the medial side tended to be not so well developed as their counter- parts on the lateral side and thus the basic pattern, although very similar, was often obscured. Anterior uterine artery. This artery was poorly developed and normally consisted of two or even three minor arteries of varying origin instead of the single large artery of the lateral side (Fig. 10, 2). The largest of these anterior arteries arose from the base of the hypogastric artery. It rarely reached the inferior uterine artery, instead it anastomosed with branches of this artery at the junction of the ventral and medial surfaces. Sometimes the anterior artery bifurcated to give two branches, or else a second branch arose from the superior uterine artery to run parallel with the first anterior uterine artery. In nineteen cases the main anterior uterine branch was strongly developed because it gave rise to the lateral uterine artery. However, after this latter artery branched off the anterior uterine artery reverted to the normal small size. Lateral uterine artery. The lateral uterine artery on the medial side was basically similar in origin, pattern and termination to that of the lateral side. It arose from the anterior uterine artery (nineteen cases) or from the superior uterine artery (nineteen cases). In four cases branches from both superior and anterior uterine vessels fused to form the main part of the artery. This artery was frequently poorly developed in its mid-region (fifteen cases) sometimes being less than half the diameter of its anterior and posterior ends. It gave off branches along its length both dorsally and ventrally. In all cases but one the lateral uterine artery terminated by fusion with the Medial posterior uterine artery (Fig. 10, 11). In the exception it passed around the posterio-ventral surface of the gland to fuse with the Lateral posterior uterine artery. The utero-vaginal artery occurred less frequently than on the lateral side (nineteen cases). When it was absent the blood supply to the utero-vaginal region was similar to that described for the lateral side. Posterior uterine artery. A description of the origin of this artery and its variations Blood supply of the avian oviduct 497 was given when considering the lateral side. Apart from the exceptions already mentioned its anatomy was basically the same as that of the Lateral posterior uterine artery. 13

Fig. 10. Medial side of the fowl's shell gland; normal arterial pattern. 1. Hypogastric artery; 2, anterior uterine artery; 3, inferior oviducal artery; 4, inferior uterine artery; 5, artery connecting inferior uterine artery with posterior uterine artery; 6, vaginal branch of inferior uterine artery; 7, lateral uterine artery; 8, superior uterine artery; 9, Lateral posterior uterine artery; 10, utero-vaginal artery; 11, point of junction of the lateral and posterior uterine arteries; 12, Medial posterior uterine artery; 13, middle uterine artery; 14, left internal iliac artery; 15, left pudendal artery; 16, left pelvic artery; 17, vaginal arteries; 18, superior oviducal artery.

Fig. 11. Medial side of the turkey's shell gland; normal arterial pattern. Key as in Fig. 10.

The turkey The anterior uterine artery on the medial surface was poorly developed, con- sisting of two branches of the hypogastric artery (Fig. 11, 2). The vessels supplying the medial side had a strong tendency to pass vertically on the surface. Thus the 498 Ri. D. HODGES face was mainly supplied by branches of the superior uterine artery and the inferior uterine artery which anastomosed together.

The duck In the duck the anterior uterine artery on the medial face was always composed of two or three small branches of the hypogastric artery. In two cases the Medial lateral uterine artery arose from a branch of the left internal iliac artery, resembling in its origin a middle uterine artery as found in the fowl.

The fowl The veins In nearly all instances the arteries of the shell gland were paralleled by corresponding veins. This usually occurred throughout the entire vascular tree of the gland, but certain exceptions were noted. The hypogastric vein. This vein presented a much more complex structure than the hypogastric artery. The two vessels did not normally run in parallel, as the vein drained into the renal rehevens vein on the posterior lobe of the kidney, whilst the artery arose between the middle and posterior lobes. However, in two instances the vessels did run together for most of their length, and in fifteen instances there was a small branch of the vein which paralleled the hypogastric artery for at least part of its length, supplying veins to the isthmus (Fig. 12, 2). At the point where the hypogastric vein left the shell gland it often divided to form a small plexus of veins. In eight cases this was very complex and in the remainder it varied in complexity down to a single large vein (Fig. 12, 1). Where a plexus occurred there were normally two or more points of entry into the renal rehevens vein. The superior uterine vein. The anatomy of this vein was very variable. In twenty- four cases it ran almost the full length of the superior surface of the gland and in all but two of these it eventually joined either the left internal iliac vein or one of the posterior uterine veins (Fig. 12, 7). In nineteen cases the vein was much longer than the superior uterine artery and in ten cases there was a vein but no artery present. In a further ten cases the vein and artery were the same length. The superior uterine vein terminated anteriorly at some point in the hypogastric vein complex. In those cases where the lateral uterine vein joined the superior uterine vein it drained into the latter in a corresponding manner to that found in the arteries. Posterior to the junction of the lateral uterine vein the superior uterine vein received a number of medium-sized tributaries draining the dorsal surface (Fig. 12, 7). Posterior uterine veins. These, in general, followed their corresponding arteries. However, it was noticeable that posterior uterine veins were much more frequently absent than the arteries. Thus, whereas there were seven Lateral posterior uterine arteries missing, there were nineteen cases in which the Lateral posterior uterine vein was absent (Fig. 12). In six of the former cases both vein and artery were missing. On the medial side there were only eight cases without posterior uterine veins. Considering the nineteen cases with Lateral posterior uterine veins missing, in Blood supply of the avian oviduct 499 fourteen the lateral uterine vein continued on to the vagina to anastomose with the vaginal veins. In the remaining five cases there was no connection with the vaginal veins and all the drainage was anteriorly through the lateral uterine vein.

4 5 Fig. 12. Lateral side of the fowl's shell gland; normal venous pattern. 1, Hypogastric vein; 2, vein paralleling hypogastric artery; 3, anterior uterine vein; 4, inferior oviducal vein; 5, inferior uterine vein; 6, lateral uterine vein; 7, superior uterine vein; 8, internal iliac vein; 9, pelvic vein; 10, pudendal vein; 11, Medial posterior uterine vein; 12, vaginal veins.

4 Fig. 13. Lateral side of the duck's shell gland; normal venous pattern. Key as in Fig. 12. The turkey The veins paralleled the arteries exactly, with only two major exceptions. In one bird the superior uterine artery anastomosed with the pelvic artery but the superior uterine vein fused with a single, centrally placed posterior uterine vein. In the second bird the superior uterine artery did not run the full length of the dorsal surface. The superior uterine vein, however, did so and, dividing towards its termination, each half fused with the posterior uterine vein of its own side. 500 R. D. HODGES

The duck The hypogastric vein consisted of a single large vein and did not form a plexus as in the fowl (Fig. 13). The general pattern of the veins followed that of the arteries with two exceptions: first, the lateral uterine veins branched from a superior uterine vein which ended at its division into these veins. Secondly, there were no posterior uterine veins; the lateral uterine veins ended on the utero-vaginal junction.

The microscopic anatomy of the shell gland vessels in the fowl It has been mentioned previously that the blood vessels supplying the oviduct were arranged in the form of an anastomotic plexus, arteries interconnecting with arteries and veins with veins. This was seen in the major vessels supplying the

Fig. 14. Surface view of the arterial supply to a small area of the fowl's shell gland (latex cast; approximately x 3). 2, anterior uterine artery; 7, lateral uterine artery; 8, superior uterine artery; PB, primary branches of the main arteries; SB, secondary arterial branches; TB, tertiary arterial branches. oviduct (Figs. 2, 3) and in the main vessels of the shell gland (Figs. 6, 12). This plexiform pattern also occurred right down to the smallest vessels in the shell gland tissue (Figs. 14, 15). The wall of the shell gland is composed of seven layers of tissue (see Fig. 16) and the inner, glandular layer is arranged in flat, leaf-shaped folds. The blood vessels which were seen externally in an injected specimen lay in the connective tissue between the two muscle coats. Fig. 14 is a diagram of the arteries in a small area Blood supply of the avian oviduct 501 of the dorso-lateral surface of shell gland no. 33. It shows the largest vessels (marked 2, 7 and 8) and their primary branches (PB). The secondary branches (SB) mainly passed out of sight through the circular muscle coat into the submucosa. The specimen was then macerated and a small portion, shown by the dotted line in Fig. 14, was closely examined (Fig. 15). The largest vessels were the primary arterial branches. The secondary branches often passed underneath them, indicating their passage into an inner tissue layer. The secondary branches divided up to form the tertiary branches (TB) lying above the glandular layer. Many of these branches terminated in the capillary beds of the glandular folds (represented in Fig. 15 by the dotted areas).

Fig. 15. Magnified view of sector of Fig. 14 after maceration, showing the blood supply to the secretary epithelium (approximately x 14). Key as in Fig. 14. Examination of the indian ink injected specimens showed that the muscle and connective tissue layers were supplied from the particular arterial branches of their own level. Throughout the gland the veins showed a very similar pattern to that described for the arteries. An overall picture of the blood supply to the shell gland wall can be seen in Fig. 16. This is part of a cleared, ink-injected specimen. It shows the main vessels lying in the first connective tissue layer, the smaller vessels lying in the submucosa (secondary and tertiary branches) and the fine capillary networks of the glandular layers. Fig. 16 shows the injected capillary network of a single glandular fold. The larger vessels running up through the connective tissue corium of the fold are tertiary branches (arterioles), supplied from the vessels in the submucosa. In the corium they divide into capillary branches which pass outwards between the 502 R. D. HODGES

I. 4n'' -

16 Blood supply of the avian oviduct 503 tubular glands, and these then form a capillary network underneath the internal ciliated epithelium. The largest, central vessel is a venule draining the glandular fold. The complete blood supply to the shell gland wall is illustrated diagrammatically in Fig. 18. Shell gland activity An attempt was made to correlate the findings in the present material with the state of activity of the shell glands by comparing the weights of the latex casts. The latex was injected at normal blood pressure in order to give a cast of the blood vessels as close as possible in volume to the original vascular volume of the shell glands. The macerated latex casts were blotted dry with a cloth and then rapidly weighed before any absorbed moisture could evaporate. Groups of fourteen active and seven quiescent glands, all well injected, were compared and the mean values and standard deviations found were: active glands 1-85 g. ± 034; quiescent glands 1-08 g. + 021. When subjected to the 't' test these figures showed significance at the 5 0 level. Thus if it can be assumed that there is a correlation between the weight of the latex cast and the vascular volume ofthe shell gland, there is on average an increase of 42 0 between the vascular volume of an active gland over a resting gland.

DISCUSSION This account of the shell gland vasculature of the fowl corresponds fairly closely with those of Freedman (1962) and Freedman & Sturkie (1963). The main difference occurs in the designation of posterior and middle uterine vessels. For reasons already given it has been decided to call the vessels which supply the posterior third of the gland the posterior uterine vessels and to reserve the designation middle uterine vessel for a vessel with a similar point of origin but a more anterior and dorsal termination. This was because with the wide but continuous range of variation of posterior uterine vessels found in this series, it was not possible to divide them at any point into middle and posterior uterine vessels. Examination of fig. 2 in Freedman & Sturkie (1963) shows a middle uterine artery which would have been considered here to be a posterior uterine artery at the anterior end of the range of variation. These differences in nomenclature probably depend at least partly upon the anatomical variations which may occur between the different breeds of fowl which were used. This is supported by the high degree of variation which was observed here. In their accounts Freedman & Sturkie gave little attention to the medial side of the shell gland, which does vary in its vasculature from the lateral side, and they did not mention the inferior oviducal artery and vein which join with the Lateral anterior uterine vessels on the ventral surface of the gland. Fig. 16. Thick section through the ink injected, cleared wall of an active shell gland, showing the disposition of the different parts of the blood supply ( x 12-5). 1, leaf-shaped folds of the secretary epithelium (side view); 2, longitudinal and circular muscle layers with intermediate connective tissue; 3, primary blood vessel branches; 4, secondary vessel branches passing through the circular muscle layer; 5, secondary vessel branches in the submucosa; 6, tertiary blood vessel branch passing into a secretary fold. Fig. 17. Single leaf-shaped fold of the secretary epitheliumn of the shell gland injected and cleared to show the ventral blood vessels (v) and their branches passing outwards to supply the peripheral capillary network (n) ( x 70). 504 R. D. HODGES There were very few significant differences between the three species studied. In the case of the oviduct the main difference was between the fowl, which normally possessed anterior oviducal arteries and veins, and the turkey and duck which did not. In the case of the shell gland the hypogastric-anterior uterine-inferior uterine- inferior oviducal vessel complex was very similar in all three birds. Differences were found in the posterior uterine vessels, but as the interspecific variations were often also found within the group of fowls, not too much importance can be attached to them. The greatest difference was found in the lateral uterine vessels. In the fowl these were partly or wholly vessels passing horizontally back along the mid-line

Fig. 18. D)iagrammatic representation of a cross-section of the shell gland wall demon- strating the blood supply. 1, Peritoneum; 2, longitudinal muscle layer; 3, connective tissue containing main vessels and primary branches; 4, circular muscle layer; 5, the submucosa containing secondary branches; 6, tertiary vessel branches in the corium; 7, tubular gland layer; 8, ciliated epithelium; 9, capillary network.

of the lateral or medial faces of the gland, whilst in the turkey and duck they ran diagonally back across the dorso-lateral aspect of the glands between the superior and posterior uterine vessels. It was noted in the latex casts of the quiescent glands that the capillary beds of the epithelial folds were hardly ever filled by the latex, and this observation was confirmed in the ink injected glands, the ink rarely penetrating further than the tertiary arterial branches entering the glandular folds. Sections taken from the latter glands showed that the epithelial capillaries still contained erythrocytes, indicating that the preliminary saline injection had failed to enter these vessels. Examination of sections of ink-injected active glands showed that these capillary Blood supply of the avian oviduct 505 beds, although often incompletely filled by the ink, had been entirely washed free of blood cells by the saline. Attempts were made to measure capillary diameters of active and quiescent glands but, although satisfactory results were obtained for ink filled capillaries, it was not possible to obtain accurate measurements of vessels still containing blood. It would seem from these results that in the quiescent state the blood flow through the secretary epithelium is considerably reduced, probably by contraction of the arterioles supplying the glandular folds. Although there is little or no flow through the capillary beds during injections of quiescent glands, the fact that the ink flows easily through the gland indicates the possible presence of arterio-venous anastomoses at the level of the secondary blood vessel branches in the submucosa. How- ever, no such anastomoses have been found in the casts. There are a number of errors inherent in the use of the latex casts to estimate changes in the vascular volume of the shell gland. First, latex has a tendency to shrink on setting. Secondly, the degree of correlation between the weight of the latex casts and the in vivo volume of the vessels which they represent is very uncertain. Thirdly, the fact that the capillary beds in the quiescent glands are not injected by the latex brings in an error of unknown size. However, it was hoped that these disadvantages might tend to cancel each other out when comparisons were being made between glands and it is probable that the results do indicate the degree of variation which occurs in the vascular volume between the active and (uiescent states. S U A A AR Y 1. Fowls, ducks and turkeys were injected with Neoprene latex or indian ink in order to demonstrate the anatomy of the blood supply to the oviduct and in particular the shell gland. 2. In all three species there was an arterial supply from the internal iliac and sciatic vessels. The fowl also had a vessel supplying the anterior oviduct from the ovarian artery whilst in the turkey and duck this was replaced by a branch of the femoral artery. 3. The venous drainage of the oviduct was mainly through the internal iliac and renal rehevens veins in all species. 4. A more detailed examination and description was made of the shell gland blood supply, both lateral and medial sides. The arteries supplying the gland were hypogastric, left internal iliac and inferior oviducal arteries and the veins through which drainage occurred were the corresponding veins lying parallel to these arteries, except in the duck where there was no drainage into the internal iliac vein. 5. Finally, a study was made of the microscopic anatomy of the blood supply to the shell gland wall of the fowl. The author wishes to express his thanks to Dr A. H. Sykes for his interest and advice, to Miss C. Freeman for technical assistance and to Mr B. J. Emmett for the photographic illustrations. 506 R. D. HODGES

REFERENCES BAntKOW, H. (1829). Anatomisch-physiologische Untersuchungen vorzuiglich fiber das Sclhagader- system der Vogel. lrch. Anat. Plhysiol. 4, 305-496. BiRADLEY, 0. C. & GRIAIIAMIE, T. (1960). The Structure of the Fowl (4th ed.). London: Oliver and Boyd. FREEDMAN, S. L. (1962). Innervation and blood vessels to the uterus of the chicken. Doctoral dissertation, Rutger's University, N. J. FREEDMAN, S. L. & STURKIE, P. D. (1963). Blood vessels of the chicken's uterus (shell gland). Am. J. Anat. 113, 1-7. HUNSAKER, AV. G. (1959). Blood flow and calcium transfer through the uterus of the chicken. Doctoral thesis, Rutger's University, N. J. MAUGER, H. MI. (1941). The autonomic innervation of the female genitalia in the domestic fowl and its correlation with the aortic branchlings. Am. J. vet. Res. 2, 447-452. NEUGEBAUEIR, L. A. (1845). Systema venosum avium cum eo mammalium et imprimis hominis collatum. Nova Acta Academiae Leopoldino-Carolinae Naturae Curiosorum, Band 21, Teil 2. OI{SINI, MARGARET AV. (1962). Technique of preparation, study and photography of Benzvl Benzoate cleared material for embryological studies. .1. Reprod. Fert. 3, 283-287. OwrEN, R. (1866). On the Anatomy of Vertebrates, vol. II, Birds and Mammals. London: Longmans, Green. RICHARDSON, K. C. (1935). The secretary phenomena in the oviduct of the fowl, including the processes of shell formation examined by the iicroincineration technique. Phil. Trans. R. Soc., Ser. B, 225, 149-195. STRESEIMAN, E. (1928). Ilandbuch der Zoologie, vol. III, Part 2. Berlin: Walter de Gruvter. TONIPSETT, 1). H. (1956). Anatomical Techniques. London: Livingstone.